Move VZEXT_MOVL nodes up through shift nodes. We should be trying harder to move VZEXT_MOVL towards any associated SCALAR_TO_VECTOR nodes to make use of MOVD/Q implicit zeroing of upper elements. Fixes #141475
51 lines
1.9 KiB
LLVM
51 lines
1.9 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
|
|
; RUN: llc -mtriple=x86_64-unknown-linux-gnu < %s | FileCheck %s
|
|
|
|
; Make sure codegen doesn't try to inspect the use list of constants
|
|
|
|
; Make sure we do not try to make use of the uselist of a constant
|
|
; null when looking for the alignment of the pointer.
|
|
define <2 x i32> @no_uselist_null_isDereferenceableAndAlignedPointer(i1 %arg0, ptr align(4) %arg) {
|
|
; CHECK-LABEL: no_uselist_null_isDereferenceableAndAlignedPointer:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: xorl %eax, %eax
|
|
; CHECK-NEXT: testb $1, %dil
|
|
; CHECK-NEXT: cmoveq %rsi, %rax
|
|
; CHECK-NEXT: movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
|
|
; CHECK-NEXT: retq
|
|
%select.ptr = select i1 %arg0, ptr null, ptr %arg
|
|
%load = load i32, ptr %select.ptr
|
|
%insert = insertelement <2 x i32> zeroinitializer, i32 %load, i64 0
|
|
ret <2 x i32> %insert
|
|
}
|
|
|
|
; Make sure we do not try to inspect the uselist of a constant null
|
|
; when processing a memcpy
|
|
define void @gep_nullptr_no_inspect_uselist(ptr %arg) {
|
|
; CHECK-LABEL: gep_nullptr_no_inspect_uselist:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: movzbl 16, %eax
|
|
; CHECK-NEXT: movb %al, (%rdi)
|
|
; CHECK-NEXT: retq
|
|
%null_gep = getelementptr i8, ptr null, i64 16
|
|
call void @llvm.memcpy.p0.p0.i64(ptr %arg, ptr %null_gep, i64 1, i1 false)
|
|
ret void
|
|
}
|
|
|
|
define <16 x i8> @load_null_offset() {
|
|
; CHECK-LABEL: load_null_offset:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: movzbl 11, %eax
|
|
; CHECK-NEXT: movd %eax, %xmm0
|
|
; CHECK-NEXT: pslld $8, %xmm0
|
|
; CHECK-NEXT: retq
|
|
%gep.null = getelementptr i8, ptr null, i64 11
|
|
%load = load i8, ptr %gep.null, align 1
|
|
%insert = insertelement <16 x i8> zeroinitializer, i8 %load, i64 1
|
|
ret <16 x i8> %insert
|
|
}
|
|
|
|
declare void @llvm.memcpy.p0.p0.i64(ptr noalias writeonly captures(none), ptr noalias readonly captures(none), i64, i1 immarg) #0
|
|
|
|
attributes #0 = { nocallback nofree nounwind willreturn memory(argmem: readwrite) }
|